Non-profit platform that catalogs, archives, and connects African language technology resources (datasets, papers, models, tools). Addresses the discoverability problem -- African language resources are fragmented across many platforms and hard to find. Covers 2,199 African languages with 1,000+ AI resources catalogued. Uses custom algorithms to detect which African language(s) a resource involves, even when not explicitly stated. Participatory approach allows community contributions. Goal: 200,000 linked records by 2026. Founded by Chris Emezue (Nigeria) and Bonaventure Dossou (Benin), both core Masakhane researchers. Advised by Turing Award winner Yoshua Bengio. Won first place at UNESCO/ViVaTech 2021 challenge for "Overcoming Language Barriers through Data and AI."
Essential discovery platform for African language technology resources. Any UNICEF programme seeking to identify available NLP tools, datasets, and models for WCA languages should use Lanfrica as a starting point. The founders are deeply embedded in the Masakhane and broader AfricaNLP community. Could serve as a resource mapping tool for UNICEF's technology selection process. Won UNESCO/ViVaTech challenge.
Lanfrica is a meta-resource (catalog) rather than a technology builder. The founders also run NaijaVoices (Chris Emezue) and contributed to Masakhane/FFR (Bonaventure Dossou). Yoshua Bengio's advisory role adds credibility. Platform code is not open source but the catalog itself is freely accessible.